άμε and άντε: Semantics Persistence in Modern Greek hortative particles

By: | Post date: 2011-03-01 | Comments: 3 Comments
Posted in categories: Linguistics, Modern Greek
Tags: , , ,

Vasilis Orfanos wrote a magisterial post on αμέτι μουχαμέτι, which I alluded to last post, and which am using to base the next post on. (Or maybe the post after that.) To acknowledge my debt, this is a post on a topic he requested from me a few weeks back. The topic in turn is in response to an earlier post I wrote on the etymology of α σιχτίρ, which involves a particle meaning “go!”.

The post is about two of the four particles of Modern Greek meaning “go!”, άμε and ά(ι)ντε. (άντε is the Standard form, but άιντε appears to be older.) Orfanos thinks that he has identified a nuance distinguishing the meaning of the two—particularly in older texts. I have the subjective impression that there is a different (though related) nuance distinguishing between them. I’m interested in hearing whether other Greek-speakers pick up on those nuances; and I’ll propose (in my typical irresponsible way) that those nuances can be explained through persistence.

Persistence is a property that has been claimed of grammaticalisation—the process where content words turn into function words. But persistence can be said to apply in semantic change in general, and it’s not a particularly surprising concept. The claim of persistence is that, when a word changes function and starts playing a new role, that some components of the meaning of the old word persist, and are inherited by the word in its new role.

To take one example: the going to/gonna future started off with a verb indicating motion towards; the will future started off with a verb indicating desire; the shall future started off with a verb indicating obligation. The notion of moving towards a goal with going to in its new function as a future marker—as Bybee & Pagliuca have argued; for example, in its stronger emphasis of preexisting intention, and on the apparent inevitability of the future event, in contrast to will futures.

Bybee, Joan L. & Pagliuca, W. 1987. The evolution of future meaning. Papers from the 7th International Conference on Historical Linguistics. Amsterdam: John Benjamins. 247–264. This paper by Steve Nicolle has a couple of examples from them towards the end.

To take a more concrete example: in many languages, a verb for “take” ends up being an accusative marker: “he takes a stick, breaks (it)” ends up interpreted as “he, (accusative) a stick, breaks”; that has happened in the African language Twi, with the verb-turned-particle de.

Example from Lord, Carol. 1993. Historical change in serial verb constructions. Amsterdam: John Benjamins.

Now, de has a seemingly arbitrary restriction, that you can only use it before objects that already exist: you cannot use if for an object that the verb brings into being. You can say “he de fire brought to his brother”, but you can’t say “he de letter wrote to me”. With persistence, though, the restriction makes sense: you can only take things if they already exist. You can take a fire and bring it to your brother, but you can’t take a letter and *then* write it to your brother. de in its new function is merely continuing to precede the kinds of objects that “take” precedes.

That’s our linguistic theory; let’s see how we can apply it. Our two particles, άμε and άντε, look like imperatives. They look so much like imperatives, that they have plural counterparts: άμε “you (singular) go!” has a plural αμέτε—which we’re about to run into again in connection with αμέτι Μουχαμέτι. άιντε also has a plural άιντεστε. Both plurals are slightly odd: the regular plurals should have been άμετε (which apparently exists, though I’ve never heard it), and *άντετε. I can guess why they look like they do, but the plurals aren’t the point of the article.

(For the record though, my guess is: *άγωμε-τε > *άγωμέτε > αμέτε, and άντε > άντες > άντεστε by analogy with βγες > βγέστε. I see the Triandaphyllides Institute dictionary thinks instead αμέτε is accented by analogy with ελάτε “come!”; *shrug* maybe.)

In fact, the plurals are Modern Greek linguistic creativity: the two words άμε and άντε look like singular imperatives (they end in -e), and they act like imperatives (πήγαινε στο πηγάδι “go to the well”, άμε/άντε στο πηγάδι “go! to the well”). But historically they are not imperatives at all.

άμε originates in the Ancient 1st person plural subjunctive ἄγωμεν, “let us lead!” = “let us go!”. (The switch from “lead” to “go” for ἄγωμεν in particular is already in the New Testament: Mark 1:38 “Let us go somewhere else—to the nearby villages”, ἄγωμεν ἀλλαχοῦ εἰς τὰς ἐχομένας κωμοπόλεις.) In Early Modern Greek, a lot of compulsory final /n/’s got treated as optional, liaison /n/’s. That meant ἄγωμεν ended up as άγωμε, which now looked like an imperative, and could be applied to “you” rather than “us”. But that’s the final stage in a development that must already been underway for a long time beforehand.

The same development is underway right now for let’s in English. Historically let’s is let us, so it only refers to an action undertaken by us. But let’s is gradually starting to be used as if the “us” is not there: let’s you and him fight may be used in jest, but the seemingly redundant let’s you and I (270,000 Google hits) clearly isn’t. What’s happening with let’s is classic semantic bleaching: the “us” component of the meaning of let’s is starting to fade away; the “this should happen” component remains.

The same looks like it has happened for ἄγωμεν: the meaning has shifted from the Classical “let us lead!”, to Jesus’ “let’s go!”, to a subsequent bleached “someone should go”, to a refocused “*you* should go”. The switch is certainly complete by ca. 1364, and the Entertaining Tale of Quadrupeds, where the horse says to the donkey: ἄγωμε μὲ τὴν κάμηλον τὴν μακροσφονδυλάτην, “go with the long-necked camel!” (and certainly not “let’s go with the long-necked camel”: the point of the poem is pairs of animals argue about their merits, and one drives the other off the stage).

In fact, άγωμε is pretty out of place as a vernacular verb in 1364. The verb άγω is not supposed to have survived, and a distinct present subjunctive (άγωμε as distinct from the indicative *άγουμε) certainly should not have survived. That this subjunctive had survived this long indicates that people no longer realised it was a subjunctive at all: it was by then just a particle, like let’s is starting to be. And as a particle, άμε was ripe for reanalysis as an imperative.

That’s άμε. The other particle, άιντε ~ άντε, has been derived from the same verb, tortuously, as ἄγε δή, “go indeed!” or ἄγετε “you all, go!” The proposals doesn’t make phonological sense, and δή looks particularly suspicious; but they serves a greater political goal—and Brian Joseph has already lambasted them for it:

Joseph, B.D. 1985. European Hellenism and Greek Nationalism: Some Effects of Ethnocentrism on Greek Linguistic Scholarship. Journal of Modern Greek Studies 3: 87–96. p. 91.

A Hellenic etymology of άιντε avoids the blindingly obvious etymology of άιντε, which is Turkic: Turkish haydı.

(The Triantaphyllides Institute etymology is from άμετε > άμτε > άντε. Less insane, but I don’t particularly see the plural origin as necessary; and to their credit, they also allow the derivation I think obvious.)

Forms of hayda “go!” turn up not only in Turkish hayda ~ haydı, but also Crimean Turkish hayde, and Tatar aida “interjection of encouragement”, eyde “listen! faster! go ahead!” They also turn up throughout the Balkans—as well as Russian and Czech, thanks to the Tatars and not the Greeks: Russian (г)айда (g)ajda, Czech hajdy.

The etymology of haydı is not settled. If it breaks down as hay + da, (h)ay could be either the interjection (h)ay “ah!”, associated with compassion, terror, nostalgia or a call to attention; de would then be the particle de, expressing impatience (which Modern Greek knows as ντε, and which corresponds to Yiddish-flavoured alright already). But while hay de would explain haydı through vowel harmony, it wouldn’t explain hayda.

Less problematically, haydı could originate from the verb ayda “to drive animals; to steal animals away from their owner”. That verb has a much further spread into Asia, all the way to Tuva (Chagatai and Uighur haida, Tuvan aida, Tatar ayda, eyde, Crimean Turkish aida, Turkish hayda, Kazakh aida, Kirghiz aida). Not so coincidentally, that verb has the same meaning that ἄγω originally had. And the verb may ultimately be related to the interjection ay after all: ayda may come from “ay” la, “to shout ‘hey!'”.

Turkic data and etymologies from: Radloff, W. 1893–98. Опыть Словаря Тюркскихъ нарѣчій. St Petersburg: Glasunoff.

Those are our etymologies: άμε from “let’s go”, a exhortion—and, more remotely, from a verb for driving cattle; άντε directly from a verb for driving cattle (and possibly influenced by a particle expressing impatience). Both were tied up three thousand years ago with driving cattle, but their more recent past, and their distinct reference to “us” vs “you”, should differentiate them. That’s the point I was making last post about metaphorical accounts of language change: what the word meant 3000 years ago is not as important for persistence as what the word meant 1500 years ago.

So given persistence, we *could* expect that άμε should be more about encouragement, and άντε more peremptory: άμε as a more positive notion, and άντε as a more negative. FWIW, I get these Google counts for a common blessing and curse with the two particles:




στο καλό “to the Good”




στο διάολο “to the Devil”




Not an overwhelming difference, but it is something.

The distinction Vasilis Orfanos has observed from older Modern Greek texts and folksong is is different:

When we urge someone to leave from the place we are at and to go somewhere else, we have an itinerary in mind. From the examples I have gathered, mainly from older Greek texts and folk song, it seems to me that when we use άμε we are imagining them arriving at the end of the itinerary, while when we use άντε it’s their departure that dominates in our mind. Of course as the years passed, the one usage influenced the other.

In other words, άμε emphasises the destination of movement, and άντε the origin of movement. In more crude terms (and not necessarily quite the nuance Orfanos is appealing to), άμε implies “get to X”, άντε implies “go away from X”.

If that distinction was in place, we can tell a story to explain it: given the egocentric perspective of language-speakers (which is just cognitive psychology, and not a character flaw), we want good things to happen to good people, but we want bad things to be away from us. If we’re sending off both a good thing and a bad thing, we don’t care where the bad thing ends up, as long as it’s the hell away from us: that’s the emphasis on the origin of movement, on “go away”. The good thing, on the other hand, we want to succeed; and success is arriving where it set out to go: that’s the emphasis on destination, on “get to X”.

I’m not overwhelmingly convinced by my argument, but it’s possible. A problem is, while άμε is limited to encouragement of movement (persistence yet again, since ἄγω was a verb of movement), άντε is used quite broadly, and can also be as a word of (pleading) encouragement: “oh, go on!”.

(The range of meanings άντε can take is extraordinary, all the way to illocutionary exhortation: έξι, άντε εφτά “six, at most seven” is actually literally “six—oh, go on then, seven”, meaning “go on, have it your way, I’ll allow the number’s closer to what you think than what I think”. How the pragmatics works is another post. How closely the Greek range of άντε matches the Turkish range of haydı is another post still.)

I do still think that, when used specifically with respect to movement, and without a whining tone of voice, άντε is by default negative, “get away from here”; pleading, after all, is still closer to “I wish you’d go away” than “I hope you arrive safely.” So I don’t think the pleading sense of άντε is inconsistent with an emphasis of movement away from X, rather than towards X.

But that’s me. What do you think? If you speak Greek, do either of our surmises on the connotations of άμε and άντε match your understanding of Greek?


  • gbaloglou says:

    When a non-Greek word gradually displaces a Greek word (in Greek texts) … shouldn't we expect the displacement(s) to evolve from syntactically simple situations to complicated ones? If this postulate holds true, and if we admit that άντε gradually displaced άμε, then perhaps we have another explanation for άντε's emphasis on departing point: "get out of HERE" is simpler, I submit, than "go to such and such place" — "here" is obvious, "there" is more abstract and demands some explanation, however minimal.

  • gbaloglou says:

    I am surprised to find eight examples — in fact one, it's a song — of "άντε στο καλό κι η Παναγιά μαζί σου" and one example of "άντε στο καλό κι η Παναγιά μαζί σας", but none of "άμε στο καλό κι η Παναγιά μαζί σου/σας": that's in google, I do not know about older texts.

  • John Cowan says:

    I think you're jumping the gun on let's. As long as the serious example is let's you and I, then since us includes (or may include) both you and I, there is no real grammaticalization. Let's you and I go is not only the syntactic descendant of "Let us go now, you and I / When the evening is spread out against the sky", but crucially also has the same semantics.

    When we start to see let's you go and let's them go in ordinary discourse, then it's time to talk of grammaticalization.

Leave a Reply

  • Subscribe to Blog via Email

    Join 297 other subscribers

  • December 2017
    M T W T F S S
    « Nov    
%d bloggers like this: